1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Scythe Infinity: Dual Fan vs. Single Fan

Discussion in 'Overclocking & Cooling' started by s1rrah, Jan 25, 2007.

  1. s1rrah New Member

    Joined:
    Oct 21, 2006
    Messages:
    263 (0.09/day)
    Thanks Received:
    3
    (i've posted this on other forums, but just noticed i never posted it here; don't know how many techpowerup! users are on other forums, so thought i'd post it here as well)


    scythe infinity: single fan vs. dual fan
    --------------------------------------------------------------------

    as discussed in several other threads, i ordered some ducting from thermalright along with a second silverstone FM121 to mount with my already installed Scythe Infinity and originally purchased, identical FM121 fan.

    as it turns out and after a bit of installation testing ... it ends up that the ducting won't work because the heat-sink end of the duct was about 1.5" lower than the case exhaust portal and the particular thermalright ducting i had purchased had nearly zero vertical flexibility. with ducting that had equal 360 degree flexibility, it would work ... but not so with what i ordered. this of course is no fault of thermalrights ... this is, afterall a hack job of sorts.

    here's an illustration of the ducting problem:

    ...


    the ducting problem:

    [​IMG]

    ...

    so, though bummed, I still mounted the second 110 CFM silverstone alongside the first, in a push/pull configuration ... so as to determine what sort of performance increase could be gained by using two fans; also, i was/am interested in maximum cooling with the lowest sound for 24/7 usage at a fairly high overclock of 3.6ghz on my 2.1ghz E6400.


    the main cooling goal
    ------------------------------------------------------------------

    the main goal is twofold:

    1. fairly big 24/7 overclock on a E6400 with the lowest sound possible.
    2. when very cold room ambients are available (depending on weather) i want max heat dissipation without any regard for sound, simply for the purposes of benchmarking and maybe melting a component or two.

    keep that in mind when reading on ... because for benching sake and when running both silverstone's at max RPM's ... schwew! ... things get pretty loud.


    fan's used
    ------------------------------------------------------------------

    2 x 110CFM Silverstone FM121's

    These fan's will be mounted on either end of the scythe infinity cooler in a push/pull configuration. i'll be measuring idle and load temps at the fan's max concurrent speed of 2400RPM's as well as readings at a lower setting of 1200RPM's or so.


    zip ties (necessary evils)
    ------------------------------------------------------------------

    since scythe only shipped me one set of mounting wires for a single fan, so i had to improvise and use zip ties. furthermore, since my zip ties wouldn't fit all the way around the cooler, i had to hook two together by inserting the male end of one into the female end of a second, thereby forming one *really* long zip tie.

    the dual fan mount via this method is quite solid and secure and it's unlikely that the strap's infront of each fan are hindering the air flow in anyway.

    also, i mounted the push-fan a bit lower than the pull-fan because it would more directly hit the NB and NB-heat pipe in that configuration; i lost like a degree or so on the CPU temps, but because the NB was heating up quite a bit, i thought it would be a decent compromise:

    ...

    [​IMG]



    up and running / test data

    ------------------------------------------------------------------

    after finishing mounting the dual fans, i turned the computer on and began the series of tests:

    ...

    [​IMG]

    ...

    *** all tests conducted at room ambient of 22C (+/- .05C) ***
    *** all tests run on E6400 at 3.6ghz ***


    first i tested the dual fans at 2400RPM's, then at 1200RPM's; between each test, the system was powered down for thirty minutes with the case door open.

    idle temps ... obtained after 20 minutes system inactivity.
    load temps ... obtained after 30 minutes orthos (in-place large FFT's)

    and after completing a full suite of tests for both max and low RPM dual and single fan speeds, here's what i found:

    ...


    SINGLE fan at 2400RPM's (max):
    -------------------------------------------------------------------
    idle: 38C
    load: 54C

    DUAL fan at 2400RPM's (max):
    -------------------------------------------------------------------
    idle: 36C
    load: 52C

    ....


    SINGLE fan at 1200RPM's:
    -------------------------------------------------------------------
    idle: 39C
    load: 55C

    DUAL fan at 1200RPM's:
    -------------------------------------------------------------------
    idle: 36C
    load: 53C


    ...


    Final Thoughts...
    --------------------------------------------------------------------

    across the board, at whatever speed, the two fan's in a push pull set up reduced cpu temps by around 2 degrees on average.

    the main thing to note, however, is that the two fan's running at a low RPM setting, still outperformed the single fan at it's max RPM setting; and the two fan's running at 1200RPM were significantly, even extremely more quiet than the single fan running at max RPM's. both fan's running at 1200RPM's were near imperceptible to the ear ... i had to kill the rest of my case fan's to hear them.

    so all and all, i'm happy with the results; now i'm running even cooler and i can't even hear the heat sink fans.

    another thing i'm interested in seeing is just how much heat this dual fan set up can dissipate once i can access room ambients appropriate for 4.1ghz+ benchmarking. keep in mind that all temps were obtained with a moderately low room ambient of 22 degrees and since air benchmark's are *totally* dependant on very low room ambients ... it will be interesting to see just how much of a difference 12 degree celsius ambients will make when trying to run benchmarks at 100% overclocks or higher. :)

    hope this info is helpful.

    happy new year!
     

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page